this four-year longitudinal prospective study will utilize findings on drug-user social networks and drug use sites to test a peer-led intervention with active drug users and their networks, conducted in the places they gather to use drugs. The project is based on a study of high-risk drug use settings for HIV prevention (P01 MH/DA5 6826, Project #3). Aims of the new study, entitled "HIV Prevention in High-Risk Drug Use Sites: Project RAP (Risk Avoidance Partnership)" are to: 1) develop and implement in high-risk drug-use sites an intervention program using trained Peer Health Advocates (PHAs) who will reach site users and their risk networks with HIV prevention messages, materials, and model practices; 2) evaluate the efficacy of this site-based peer intervention program for reducing HIV risk and increasing protective behaviors among drug users and their risk network members; 3) document the diffusion of intervention messages and materials by PHAs into sites and through the drug-using networks associated with them; 4) describe the characteristics and network relationships of effective PHAs who successfully bring about long-term adoption of HIV prevention practices among their network contacts; and 5) initiate and document the development process of and activities associated with a drug-user-led Community Advocacy Group, as a vehicle for their own organization and mobilization to advocate for group-identified health concerns and other issues affecting drug users at risk. We will recruit and train 120 active drug users to conduct a specifically designed prevention intervention with their drug-using network members and sex partners. PHA training will include a 5-session curriculum followed by 5 sessions in which the PHA disseminates the intervention in the field in partnership with project staff. PHAs will refer into the project 360 network members (Contact Referrals-CRs) to whom they will provide intervention, and whom we will assess for number and type of intervention contacts, impact, and long-term effects. We also will conduct a "community-wide survey" with 750 drug users in year 4 to assess the diffusion of prevention messages/materials through Hartford drug-using networks. This study uses a panel design with PHAs and CRs to measure changes in risk behaviors and adoption of prevention practices over time (baseline, 6-month follow-up, year-4 survey). The community-wide survey additionally allows analysis of intervention effects in the broader networks of drug users. Data collection methods include structured surveys and process tracking, as well as semi-structured interviews and ethnographic observation to document project activities and effects for process and outcome evaluation analyses.